MD-SOAR Maryland Shared Open Access Repository
Not a member yet
34521 research outputs found
Sort by
Queer country: LGBTQ+ musicians are outside the spotlight as Grand Ole Opry turns 100
No full textLGBTQ+ musicians have always played a part in country music, but even today they are being kept just outside the spotlight.http://theconversation.com/queer-country-lgbtq-musicians-are-outside-the-spotlight-as-grand-ole-opry-turns-100-25189
Self-Attention Policy Optimization for Task Offloading and Resource Allocation in Low-Carbon Agricultural Consumer Electronic Devices
No full textIn recent years, the widespread use of edge agricultural consumer electronics has greatly contributed to the level of intelligence in agricultural production, bringing higher efficiency and quality. However, offloading all tasks to the cloud incurs significant latency and resource waste, while relying solely on edge computing fails to meet the computational demands of the entire system. To solve the above problems, we introduce the device-edge-cloud (DEC) three-layer architecture, where agri-consumer electronics devices can partially offload tasks to the edge, and the edge can partially offload tasks to the cloud, i.e., agri-consumer electronics can realize device-edge-cloud collaborative computation. Second, we model the joint computation offloading and resource allocation optimization problem as a non-convex optimization and propose a novel Self-Attention Policy Optimization (SAPO) algorithm to solve it. Experiments show that the joint optimization performance of the proposed SAPO exceeds the baseline, and it is suitable for many different models. Compared with fully connected networks, it has better convergence and robustness, with a convergence speed 50% faster than the fully connected networks. The proposed SAPO algorithm has good scalability and adaptability, and has the potential to be extended to smart agricultural computing scenarios with non-convex optimization.This work was supported in part by the National Natural Science Foundation of China 62462002 and partially supported by the Natural Science Foundation of Guangxi China Nos 2025GXNSFAA069958 2025GXNSFBA069394https://ieeexplore.ieee.org/abstract/document/10975106
lkspacecraft: A Tool for Obtaining Orbital Properties of the Kepler and TESS Spacecraft
No full textNASA’s Kepler and TESS missions have produced image data and time-series of millions of stars, and unlocked valuable astrophysical inference. These missions both provide precise, accurate time-series. However, because of the orbital properties of the spacecraft, the data from these missions are impacted by various relativistic effects which cause position and timing aberrations. These aberrations can be calculated using the precise ephemerides of the spacecraft, which are provided by the mission via SPICE kernels. However, SPICE kernels can be difficult for users to work with, requiring a knowledge of the workings of the SPICE package, and the location and formats of the various kernels needed to build spacecraft ephemeris. lkspacecraft is a new Python package that utilizes spiceypy to provide users with a simple interface to obtain the key properties needed to correct the relativistic effects that impact Kepler, K2 and TESS data.Funding for this work for CH, RH, NS, TP, VK, and DG is provided by the TESS Science Support Center and funded by the NASA TESS mission, under award number 80GSFC24M0006. Funding for this work for JMP and AT is provided by NASA grant 80NSSC20M0192. This package is designed to be part of the Lightkurve a Python package ecosystem for Kepler and TESS data analysis (Lightkurve Collaboration et al. 2018). This work made use of Astropy,10 a community-developed core Python package for Astronomy (Astropy Collaboration et al. 2013, 2018, 2022). This work made use of SPICE. The SPICE system software, the SPICE server and some SPICE-based tools are provided by NAIF (Acton 1996; Acton et al. 2018). This work made use of spiceypy, a Python package for working with SPICE (Annex et al. 2020).https://iopscience.iop.org/article/10.3847/2515-5172/adef3
Essays on Public Transfers and Informality
No full textThis dissertation examines key fiscal policy and labor market dynamics aspects, focusing on fiscal rules, informality, and the interaction of informal employment with social transfers. The first paper develops a countercyclical fiscal rule that addresses the limitations of traditional budget rules by incorporating countercyclicality and sustainable debt levels based on the distinct characteristics of different government expenditures. Using a macroeconomic model, it proposes a rule that enhances welfare by aligning fiscal policy more closely with optimal paths. The second paper investigates the voluntary and involuntary nature of informal work, expanding existing frameworks to account for sectoral transitions and unemployment dynamics. Empirical analysis in the context of Costa Rica reveals that a significant proportion of informal workers are involuntarily informal, with the share varying depending on assumptions about transition probabilities and mobility within the informal sector.
The third paper explores the challenges of extending unemployment benefits to informal workers, proposing conditional transfers as a solution to reduce false claims while balancing welfare gains and labor market distortions. Through a calibrated model, it evaluates how optimal transfer policies depend on informal wages, the size of the informal sector, and unemployment rates.
Together, these papers contribute to understanding how fiscal and labor market policies can be better designed to promote economic stability, reduce inequality, and enhance welfare in developing economies
Swift/XRT monitoring of the orbital and superorbital modulations in 4U 1909+07
No full textWe report on an observational campaign performed with Swift/XRT on the wind-fed supergiant X-ray binary 4U 1909+07 to investigate the nature of the orbital and superorbital modulation of its X-ray emission. A total of 137 XRT observations have been carried out, summing up to a total effective exposure time of 114 ks and covering a total of 66 orbital and 19 superorbital cycles of the source. The XRT data folded on the orbital period of the source confirmed and improved the previously reported variability in intensity and absorption column density, which can be ascribed to the neutron star accreting from the wind of its B supergiant companion across a fairly circular orbit. The XRT data folded on the superorbital period did not provide evidence of significant variations in either the absorption column density and/or the power-law photon index. This may be due to a significant weakening of the superorbital modulation during the times when the XRT observations were carried out, as confirmed by the BAT dynamic power spectrum. We discuss the implications of these findings within the corotating interaction region model proposed to interpret the superorbital variability in wind-fed supergiant X-ray binaries.This research was funded by the Programma di Ricerca Fondamentale INAF 2023 This work has been partially supported by the ASIINAF program I 004 11 5 NI acknowledges the NASA grant 80NSSC25K7622 This work was supported in part by NASA under award number 80GSFC24M0006https://www.aanda.org/articles/aa/full_html/2025/07/aa54601-25/aa54601-25.htm
I Hate the News Jul 15
No full textThe weekly news analysis from I Hate Politics: The DC Council looks to undo two citywide referendums: Initiative 82, which provided a path to equalize tipped minimum wages, and Initiative 83, which would change the district’s elections to ranked-choice voting (RCV) and holds firm. The Moore administration in MD makes an employee buy-out offer and restaffs cabinet positions as reelection year approaches and the state’s fiscal worries grow. Howard County, Maryland, considers a new approach to Adequate Public Facilities Ordinance even as school overcrowding eases. Former Congressman and billionaire David Trone endorses Even Glass for Montgomery County Executive. Music by Washington DC art-pop rock band, Catscan!https://open.spotify.com/episode/3ZZhhi9BwyRwHrmoMr4Ok
AMINO ACID ALPHABET (R)EVOLUTION: CHANGING THE MOLECULAR BASIS OF LIFE
No full textLife on Earth has evolved to construct metabolism as a network of genetically encoded proteins. Each protein comprises a sequence of amino acids, covalently linked together. Early evolution established a single, standard library of 20 L-?-amino acids with which to build proteins since life’s last universal common ancestor (LUCA). Multiple disciplinary lenses agree, however, that a far greater diversity of amino acid structures was available to life’s origins and early evolution. Amino acids appear readily available to the planetary bodies that comprise our galaxy, have been detected within meteorites, produced by a wide range of conditions for simulating prebiotic chemistry, and have even been theorized to occur within the interstellar medium. Given that a single set has proven capable of building proteins adapted to every imaginable environment for life on Earth for more than 3.5 billion years, they are an attractive, recognized focus of astrobiology research. A decade of this research now supports the idea that the genetically encoded set of 20 amino acids exhibits highly unusual physicochemical properties as a set. The range and evenness with which they cover the chemistry space of possible volume and hydrophobicity is remarkable and non-random. This theory has matured far enough that it is both tractable and useful to now ask the question: if an independent origin of life were to build proteins using amino acids then what structures and functions would we expect, using Earth life as a guide? Here I attempt to shed light on an answer to that question through the intersection of “state-of-the-art” computational and empirical methods. This work begins to uncover whether the fundamental molecular biochemistry of life elsewhere, if using monosubstituted alpha amino acids, looks eerily similar to, or starkly different from, life as we know it here on Earth. Specifically, by (i) using a heuristic search algorithm to search for and analyze alternative amino acid alphabets (ii) designing, synthesizing, and characterizing the world’s first xeno peptides; and (iii) evaluating the reliability of state-of-the-art spectral prediction algorithms for xeno amino acids. The results primarily indicate that certain amino acids are predisposed to forming high physicochemical coverage sets, the one amino acid alphabet used by life is not the only one capable of forming peptide structures, and that current spectral prediction algorithms are insufficient when simulating xeno amino acid spectra. I conclude by identifying and calling for the further development of multiple tractable research directions so as to continue uncovering what alternative biochemistries may look like
Analyzing the Sensitivity of Vision Language Models in Visual Question Answering
No full textWe can think of Visual Question Answering as a (multimodal) conversation between a human and an AI system. Here, we explore the sensitivity of Vision Language Models (VLMs) through the lens of cooperative principles of conversation proposed by Grice. Specifically, even when Grice's maxims of conversation are flouted, humans typically do not have much difficulty in understanding the conversation even though it requires more cognitive effort. Here, we study if VLMs are capable of handling violations to Grice's maxims in a manner that is similar to humans. Specifically, we add modifiers to human-crafted questions and analyze the response of VLMs to these modifiers. We use three state-of-the-art VLMs in our study, namely, GPT-4o, Claude-3.5-Sonnet and Gemini-1.5-Flash on questions from the VQA v2.0 dataset. Our initial results seem to indicate that the performance of VLMs consistently diminish with the addition of modifiers which indicates our approach as a promising direction to understand the limitations of VLMs.This research was supported by NSF award #2008812, and awards from the Gates Foundation and Adobe. The opinions, findings, and results are solely the authors’ and do not reflect those of the funding agencies.http://arxiv.org/abs/2507.2133
From Perceptions to Meaning: Multimodal and Contrastive Machine Learning to Ground Human Intentions and Language
No full textThe overarching theme of this thesis is making sense of senses—a journey from perception to meaning by teaching machines to uncover the latent connections between natural language and the physical world as perceived through multiple sensor modalities. Learning these connections is referred to as multimodal grounded language learning, and it allows AI agents and robots to interact more intuitively with their environments and communicate naturally with humans. By understanding how language aligns with the physical world, AI systems can go beyond recognizing objects and begin to infer high-level human intentions behind tasks—an essential step for building truly intelligent, interactive systems. The growing demand for personal robots as caretakers is just one example that illustrates the real-world importance of this goal. Training AI to interact effectively in such situations requires an understanding of objects through multiple sensor modalities and the language humans use to describe them. To facilitate this, I contributed to the development of a multimodal dataset of everyday objects, and proposed approaches to multimodal grounding of language, including Extended Multimodal Alignment (EMMA), a model capable of integrating any number of modalities while remaining robust to sensor failures. EMMA not only sets a new state-of-the-art by learning effectively from low amounts of data but also converges to optimal performance twice as fast. However, understanding objects is only part of the challenge—humans often communicate abstractly, omitting steps to perform a task or using indirect speech acts. For instance, when someone tells a robot, ‘the bathroom is dirty,’ the unspoken goal is ‘clean the bathroom.’ Inferring both the desired outcome and the sequence of actions required to achieve it is second nature to humans but a significant challenge for AI. To address this, I introduce Intentionality—a novel framework that grounds human intentions in tasks by learning to infer the underlying goal and the sequence of steps required to accomplish it, given natural language instructions and visual context. This thesis advances multimodal grounded language learning by enabling AI to process multiple modalities and making it more capable of interpreting human intentions and interacting intuitively with its environment
Fintech Rulemaking: Evidence from Middle East and Africa
No full textIn recent times, there has been a flurry of fintech policymaking across the Middle East and Africa (MEA) region. This chapter studies fintech rulemaking processes across the MEA region. Earlier research has described responses to fintech as (a) Wait and See, (b) Test and Learn, (c) Innovation Facilitators (includes regulatory sandboxes, innovation hubs, and accelerators), and (d) Regulatory Laws and Reform. This chapter is about the regulatory reforms with theoretical foundation in policy diffusion and administrative process. The author extends an emerging scholarship on administrative process theory of central banking. The author asserts fintech policies began with initial guidance from multilateral development banks (World Bank, IMF, UNECA, UNCDF), global standard-setting bodies (Basel Committee on Banking Supervision, Financial Stability Board), regional organization (Arab Monetary Fund), and collaborative network (Alliance for Financial Inclusion). Over time, MEA nations developed their fintech-specific regulations.The author used a self-gathered private dataset of fintech rulemaking processes across the MEA region, the World Bank Global Fintech-enabling regulations database, and the World Bank’s Global Indicators of Regulatory Governance (GRG). From the data analysis, the author’s findings led to three conceptual frameworks: (i) solicitation, (ii) engagement, and (iii) custodianship.Solicitation refers to when financial regulators (central banks, securities regulators) ask for input from the public. Engagement could take any of these form cooperation and collaboration. It refers to activities that keep regulators and the public to interact as well as help improve fintech regulatory decision-making. Lastly, custodianship explores the prerogative role held by financial regulators to guard, protect, and maintain financial inclusion, stability, integrity, and protection (I-SIP) of the financial system.https://link.springer.com/rwe/10.1007/978-981-96-6143-5_1